Compound Engineering Plugin - AI-Powered Development Workflow
Research Date: January 31, 2026 Source URL: https://github.com/EveryInc/compound-engineering-plugin npm Package: https://www.npmjs.com/package/@every-env/compound-plugin Documentation: https://everyinc.github.io/compound-engineering-plugin/ Version at Research: v2.28.0 License: MIT
Overview
The Compound Engineering Plugin is a Claude Code plugin marketplace featuring a comprehensive collection of AI agents, commands, and skills designed to make each unit of engineering work easier than the last. Developed by Every Inc., it implements a "compounding engineering" philosophy where planning and review comprise 80% of the workflow, with execution at 20%.
Core Value Proposition: Invert the traditional development cycle where features accumulate technical debt. Instead, each development cycle documents learnings, improves patterns, and makes future work easier.
Problem Addressed
| Problem | How Compound Engineering Solves It |
|---|---|
| Development work accumulates technical debt over time | Compounding cycle documents solutions and patterns for reuse |
| Code reviews catch issues too late in the development cycle | 14 specialized review agents run in parallel before code is merged |
| Knowledge silos form when only one developer knows a solution | /workflows:compound command captures solutions in searchable docs/solutions/ |
| Planning is often rushed, leading to rework | /workflows:plan includes research decision logic and Q&A refinement |
| Inconsistent code review quality across team members | Multi-agent review with domain specialists (security, performance, architecture) |
| Context loss when switching between tasks | Git worktree integration for parallel development without context switching |
| External framework documentation changes faster than team knowledge | Context7 MCP server provides real-time framework documentation lookup (100+ frameworks) |
Key Statistics (as of January 31, 2026)
| Metric | Value |
|---|---|
| GitHub Stars | 6,830 |
| Forks | 546 |
| Open Issues | 48 |
| Primary Language | TypeScript |
| Agents | 27 |
| Commands | 20 |
| Skills | 14 |
| MCP Servers | 1 |
| Contributors | 10+ |
| npm Package Version | 0.1.1 |
Core Workflow
Plan --> Work --> Review --> Compound --> Repeat
| | | |
v v v v
80% 20% 80% Knowledge
Planning Execution Review Compounding
| Command | Purpose |
|---|---|
/workflows:brainstorm |
Explore requirements and approaches before planning |
/workflows:plan |
Turn feature ideas into detailed implementation plans |
/workflows:work |
Execute plans with worktrees and task tracking |
/workflows:review |
Multi-agent code review before merging |
/workflows:compound |
Document learnings to make future work easier |
Key Features
1. Multi-Agent Review System
14 specialized review agents run in parallel:
Review Agents:
kieran-rails-reviewer- Rails code review with strict conventionskieran-python-reviewer- Python code review with strict conventionskieran-typescript-reviewer- TypeScript code review with strict conventionsdhh-rails-reviewer- Rails review from DHH's perspectivesecurity-sentinel- Security audits and vulnerability assessmentsperformance-oracle- Performance analysis and optimizationarchitecture-strategist- Architectural decisions and compliancepattern-recognition-specialist- Code patterns and anti-patternsdata-integrity-guardian- Database migrations and data integrityagent-native-reviewer- Verify features are agent-native (action + context parity)code-simplicity-reviewer- Final pass for simplicity and minimalismdata-migration-expert- Validate ID mappings match productiondeployment-verification-agent- Go/No-Go deployment checklistsjulik-frontend-races-reviewer- JavaScript/Stimulus race conditions
2. Research Agents
4 agents for gathering context and best practices:
best-practices-researcher- Gather external best practices and examplesframework-docs-researcher- Research framework documentationgit-history-analyzer- Analyze git history and code evolutionrepo-research-analyst- Research repository structure and conventions
3. Design Agents
3 agents for UI/UX implementation:
design-implementation-reviewer- Verify UI matches Figma designsdesign-iterator- Iteratively refine UI through systematic iterationsfigma-design-sync- Synchronize web implementations with Figma designs
4. Smart Planning with Research Decision Logic
The /workflows:plan command includes intelligent research decisions:
- High-risk topics (security, payments, external APIs) always trigger external research
- Strong local context (good patterns, CLAUDE.md guidance) skips external research
- Uncertainty or unfamiliar territory triggers external perspective
- Interactive Q&A refinement phase with targeted questions
5. Knowledge Compounding
The /workflows:compound command uses 7 parallel subagents:
- Context Analyzer - Extracts problem type, symptoms, YAML frontmatter
- Solution Extractor - Identifies root cause and working solution
- Related Docs Finder - Cross-references existing documentation
- Prevention Strategist - Develops prevention strategies
- Category Classifier - Determines optimal
docs/solutions/category - Documentation Writer - Assembles complete markdown file
- Specialized Agent Invocation - Post-documentation expert review
6. Utility Commands
| Command | Purpose |
|---|---|
/deepen-plan |
Enhance plans with parallel research agents |
/changelog |
Create engaging changelogs for recent merges |
/create-agent-skill |
Create or edit Claude Code skills |
/heal-skill |
Fix skill documentation issues |
/plan_review |
Multi-agent plan review in parallel |
/reproduce-bug |
Reproduce bugs using logs and console |
/resolve_parallel |
Resolve TODO comments in parallel |
/resolve_pr_parallel |
Resolve PR comments in parallel |
/test-browser |
Run browser tests on PR-affected pages |
/xcode-test |
Build and test iOS apps on simulator |
/feature-video |
Record video walkthroughs for PR descriptions |
7. Skills Library
14 skills for specialized tasks:
| Skill | Purpose |
|---|---|
agent-native-architecture |
Build AI agents using prompt-native architecture |
andrew-kane-gem-writer |
Write Ruby gems following Andrew Kane's patterns |
compound-docs |
Capture solved problems as documentation |
create-agent-skills |
Expert guidance for creating Claude Code skills |
dhh-rails-style |
Write Ruby/Rails code in DHH's 37signals style |
dspy-ruby |
Build type-safe LLM applications with DSPy.rb |
frontend-design |
Create production-grade frontend interfaces |
every-style-editor |
Review copy for Every's style guide compliance |
file-todos |
File-based todo tracking system |
git-worktree |
Manage Git worktrees for parallel development |
rclone |
Upload files to S3, Cloudflare R2, Backblaze B2 |
agent-browser |
CLI-based browser automation via agent-browser |
gemini-imagegen |
Generate and edit images using Gemini API |
brainstorming |
Guided ideation before planning |
8. MCP Server Integration
Context7 - Framework documentation lookup:
resolve-library-id- Find library ID for a framework/packageget-library-docs- Get documentation for a specific library- Supports 100+ frameworks (Rails, React, Next.js, Vue, Django, Laravel, etc.)
Technical Architecture
┌─────────────────────────────────────────────────────────────────────┐
│ Compound Engineering Plugin │
├─────────────────────────────────────────────────────────────────────┤
│ Marketplace Layer (.claude-plugin/marketplace.json) │
│ └── Plugin distribution, version management │
├─────────────────────────────────────────────────────────────────────┤
│ Plugin Core (plugins/compound-engineering/) │
│ ├── agents/ 27 specialized AI agents │
│ ├── commands/ 20 slash commands (5 workflow + 15 utility) │
│ ├── skills/ 14 skill packages with SKILL.md + scripts │
│ └── mcp-servers/ Context7 framework docs server │
├─────────────────────────────────────────────────────────────────────┤
│ CLI Converter (src/) │
│ ├── Bun/TypeScript CLI │
│ ├── Converts plugins to OpenCode format │
│ └── Converts plugins to Codex format (experimental) │
├─────────────────────────────────────────────────────────────────────┤
│ Documentation Site (docs/) │
│ └── Static HTML/CSS/JS (LaunchKit-based, no build step) │
└─────────────────────────────────────────────────────────────────────┘
Multi-Platform Support
# Claude Code (native)
/plugin marketplace add https://github.com/EveryInc/compound-engineering-plugin
/plugin install compound-engineering
# OpenCode (experimental)
bunx @every-env/compound-plugin install compound-engineering --to opencode
# Codex (experimental)
bunx @every-env/compound-plugin install compound-engineering --to codex
Installation & Usage
Basic Installation
# Add marketplace
/plugin marketplace add https://github.com/EveryInc/compound-engineering-plugin
# Install plugin
/plugin install compound-engineering
Workflow Example
# 1. Plan a feature
/workflows:plan Add user authentication with OAuth2
# 2. Execute the plan
/workflows:work
# 3. Review before merge
/workflows:review PR-123
# 4. Document learnings
/workflows:compound
MCP Server Setup (if not auto-loading)
{
"mcpServers": {
"context7": {
"type": "http",
"url": "https://mcp.context7.com/mcp"
}
}
}
Relevance to Claude Code Development
Direct Applications
- Plugin Architecture Reference: The plugin structure (agents/, commands/, skills/, mcp-servers/) demonstrates best practices for Claude Code plugin development
- Multi-Agent Orchestration: The parallel agent execution pattern in
/workflows:reviewshows effective orchestration strategies - Knowledge Management: The
docs/solutions/pattern with YAML frontmatter provides a reusable knowledge base architecture - Cross-Platform Conversion: The CLI tool for converting plugins to OpenCode/Codex demonstrates portable skill formats
Patterns Worth Adopting
- 80/20 Planning-Execution Split: Heavy investment in planning and review, minimal execution time
- Parallel Agent Execution: Run 14 review agents simultaneously for comprehensive coverage
- Smart Research Decision Logic: Context-aware research triggering based on risk and familiarity
- Compound Documentation: Structured
docs/solutions/with YAML frontmatter for searchability - Worktree Integration: Git worktrees for parallel development without context switching
- Category-Based Agent Organization: Agents grouped by domain (Review, Research, Design, Workflow, Docs)
- Interactive Q&A Refinement: Iterative clarification before plan generation
Integration Opportunities
- Agent Patterns: Review and research agent implementations could inform custom agent development
- Skill Templates: The
create-agent-skillsandskill-creatorskills provide guidance for skill authoring - MCP Server Model: Context7 integration demonstrates MCP server bundling in plugins
- Documentation Patterns: The
compound-docsskill shows structured knowledge capture - Workflow Commands: The
workflows:namespace pattern avoids collisions with built-in commands
References
- GitHub Repository: https://github.com/EveryInc/compound-engineering-plugin (accessed 2026-01-31)
- npm Package: https://www.npmjs.com/package/@every-env/compound-plugin (accessed 2026-01-31)
- Documentation Site: https://everyinc.github.io/compound-engineering-plugin/ (accessed 2026-01-31)
- Compound Engineering Philosophy: https://every.to/chain-of-thought/compound-engineering-how-every-codes-with-agents (accessed 2026-01-31)
- Story Behind Compound Engineering: https://every.to/source-code/my-ai-had-already-fixed-the-code-before-i-saw-it (accessed 2026-01-31)
- Plugin README: https://github.com/EveryInc/compound-engineering-plugin/blob/main/plugins/compound-engineering/README.md (accessed 2026-01-31)
- CHANGELOG: https://github.com/EveryInc/compound-engineering-plugin/blob/main/plugins/compound-engineering/CHANGELOG.md (accessed 2026-01-31)
Freshness Tracking
| Field | Value |
|---|---|
| Last Verified | 2026-01-31 |
| Version at Verification | v2.28.0 |
| GitHub Stars at Verification | 6,830 |
| Next Review Recommended | 2026-04-30 (3 months) |
Change Detection Indicators:
- Monitor GitHub releases for version changes (currently v2.28.0)
- Check CHANGELOG.md for new agents, commands, or skills
- Review npm package updates (@every-env/compound-plugin)
- Watch for new workflow patterns in the
/workflows:namespace - Track growth metrics (stars approaching 7K milestone)